import pandas as pd
import numpy as np
import seaborn as sns
import plotly.express as px
from matplotlib import pyplot as plt
%matplotlib inline


#html export
import plotly.io as pio
pio.renderers.default = 'notebook'

# import our data
df_2007 = pd.read_csv(r'Nairobi Securities Exchange All Stocks Prices 2007-2012\NSE_data_all_stocks_2007.csv')
df_2008 = pd.read_csv(r'Nairobi Securities Exchange All Stocks Prices 2007-2012\NSE_data_all_stocks_2008.csv')
df_2009 = pd.read_csv(r'Nairobi Securities Exchange All Stocks Prices 2007-2012\NSE_data_all_stocks_2009.csv')
df_2010 = pd.read_csv(r'Nairobi Securities Exchange All Stocks Prices 2007-2012\NSE_data_all_stocks_2010.csv')
df_2011 = pd.read_csv(r'Nairobi Securities Exchange All Stocks Prices 2007-2012\NSE_data_all_stocks_2011.csv')
df_2012 = pd.read_csv(r'Nairobi Securities Exchange All Stocks Prices 2007-2012\NSE_data_all_stocks_2012.csv')

df_2007

df_2008

df_2009

df_2010

df_2011

df_2012

# List of all dataframes
dfs = [df_2007, df_2008, df_2009, df_2010, df_2011, df_2012]

# Loop through each dataframe
for df in dfs:
    # Check if the column 'Adjust' exists
    if 'Adjust' in df.columns:
        # Rename 'Adjust' to 'Adjusted' for consistency
        df.rename(columns={'Adjust': 'Adjusted'}, inplace=True)

        # Check if the column 'Date' exists
    if 'Date' in df.columns:
        # Rename 'Date' to 'DATE' for consistency
        df.rename(columns={'Date': 'DATE'}, inplace=True)

# Combine all the dataframes from 2007 to 2012
combined_df = pd.concat([df_2007, df_2008, df_2009, df_2010, df_2011, df_2012])

# Reset the index to keep it clean
combined_df.reset_index(drop=True, inplace=True)

df = combined_df.copy()
df

# make column names uniform
df.columns = df.columns.str.lower().str.replace(' ', '_')

df.columns

Index(['date', 'code', 'name', '12m_low', '12m_high', 'day_low', 'day_high',
       'day_price', 'previous', 'change', 'change%', 'volume', 'adjusted'],
      dtype='object')

numerical_columns = ['12m_low', '12m_high', 'day_low', 'day_high', 'day_price', 'previous', 'change', 'volume', 'adjusted']

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 81504 entries, 0 to 81503
Data columns (total 13 columns):
 #   Column     Non-Null Count  Dtype 
---  ------     --------------  ----- 
 0   date       81504 non-null  object
 1   code       81504 non-null  object
 2   name       81456 non-null  object
 3   12m_low    81504 non-null  object
 4   12m_high   81504 non-null  object
 5   day_low    81504 non-null  object
 6   day_high   81504 non-null  object
 7   day_price  81504 non-null  object
 8   previous   81504 non-null  object
 9   change     81504 non-null  object
 10  change%    81504 non-null  object
 11  volume     81504 non-null  object
 12  adjusted   81504 non-null  object
dtypes: object(13)
memory usage: 8.1+ MB

# Converting specified columns

# Replace commas with empty strings in the selected columns
df[numerical_columns] = df[numerical_columns].replace({',': ''}, regex=True)

# Convert to numeric
df[numerical_columns] = df[numerical_columns].apply(pd.to_numeric, errors='coerce')

# Remove '%' symbol, handle invalid values, and convert to decimal
df['change%'] = pd.to_numeric(df['change%'].str.replace('%', '', regex=True), errors='coerce') / 100
numerical_columns.append('change%')

# Convert the 'DATE' column to datetime
df['date'] = pd.to_datetime(df['date'], format='%m/%d/%Y', errors='coerce')


df

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 81504 entries, 0 to 81503
Data columns (total 13 columns):
 #   Column     Non-Null Count  Dtype         
---  ------     --------------  -----         
 0   date       81504 non-null  datetime64[ns]
 1   code       81504 non-null  object        
 2   name       81456 non-null  object        
 3   12m_low    81504 non-null  float64       
 4   12m_high   81504 non-null  float64       
 5   day_low    81504 non-null  float64       
 6   day_high   81504 non-null  float64       
 7   day_price  81504 non-null  float64       
 8   previous   81492 non-null  float64       
 9   change     51141 non-null  float64       
 10  change%    51131 non-null  float64       
 11  volume     81504 non-null  int64         
 12  adjusted   20509 non-null  float64       
dtypes: datetime64[ns](1), float64(9), int64(1), object(2)
memory usage: 8.1+ MB

# duplicates
df.duplicated().sum()

19

#dropping duplicates
df = df.drop_duplicates()

df.duplicated().sum()

0

df.isna().sum()

date             0
code             0
name            48
12m_low          0
12m_high         0
day_low          0
day_high         0
day_price        0
previous        12
change       30344
change%      30354
volume           0
adjusted     60979
dtype: int64

# dataframe where 'previous' column has nulls
df[df['previous'].isna()]

# Drop rows where 'Previous' column is null
df = df[df['previous'].notna()]

df.isna().sum()

date             0
code             0
name            47
12m_low          0
12m_high         0
day_low          0
day_high         0
day_price        0
previous         0
change       30344
change%      30354
volume           0
adjusted     60968
dtype: int64

# dataframe where 'name' column has nulls
null_name = df[df['name'].isna()]
null_name

# getting associated codes for null names
null_name['code'].unique()

array(['OCH-R', 'NIC-R', 'DTK-R', 'HFCK-R', 'KCB-R'], dtype=object)

# dataframe with associated codes for null names
null_name_codes = df[df['code'].isin(['OCH-R', 'NIC-R', 'DTK-R', 'HFCK-R', 'KCB-R'])]
null_name_codes

# unique names for the associated codes
null_name_codes['name'].unique()

array([nan, 'Kenya Commercial Bank (Rights)',
       'Diamond Trust Bank Kenya (Rights)', 'NIC Bank (Rights)'],
      dtype=object)

# Create the dictionary for replacements
replacement_dict = {
    'NIC-R': 'NIC Bank (Rights)',
    'DTK-R': 'Diamond Trust Bank Kenya (Rights)',
    'KCB-R': 'Kenya Commercial Bank (Rights)',

    'OCH-R': 'Olympia Capital Holdings Ltd (Rights)',
    'HFCK-R': 'HF Group Ltd (Rights)'
}

# Iterate through the replacement dictionary
for code, replacement_name in replacement_dict.items():
    # Update the 'NAME' column where the 'CODE' matches and 'NAME' is null
    df.loc[(df['code'] == code) & (df['name'].isna()), 'name'] = replacement_name

df.isna().sum()

date             0
code             0
name             0
12m_low          0
12m_high         0
day_low          0
day_high         0
day_price        0
previous         0
change       30344
change%      30354
volume           0
adjusted     60968
dtype: int64

# dataframe where 'change' column has nulls
df[df['change'].isna()]

# replacing nulls with zero to indicate no change
df['change'] = df['change'].fillna(0)
df['change%'] = df['change%'].fillna(0)

df.isna().sum()

date             0
code             0
name             0
12m_low          0
12m_high         0
day_low          0
day_high         0
day_price        0
previous         0
change           0
change%          0
volume           0
adjusted     60968
dtype: int64

df['adjusted'].isna().sum()/len(df) * 100

74.83215298319689

# Drop adjusted column
del df['adjusted']

df.info()

<class 'pandas.core.frame.DataFrame'>
Index: 81473 entries, 0 to 81503
Data columns (total 12 columns):
 #   Column     Non-Null Count  Dtype         
---  ------     --------------  -----         
 0   date       81473 non-null  datetime64[ns]
 1   code       81473 non-null  object        
 2   name       81473 non-null  object        
 3   12m_low    81473 non-null  float64       
 4   12m_high   81473 non-null  float64       
 5   day_low    81473 non-null  float64       
 6   day_high   81473 non-null  float64       
 7   day_price  81473 non-null  float64       
 8   previous   81473 non-null  float64       
 9   change     81473 non-null  float64       
 10  change%    81473 non-null  float64       
 11  volume     81473 non-null  int64         
dtypes: datetime64[ns](1), float64(8), int64(1), object(2)
memory usage: 8.1+ MB

df

numericals = ['12m_low', '12m_high', 'day_low', 'day_high',
       'day_price', 'previous', 'change', 'change%', 'volume']

categorical = ['code']

# trying to see the trend of 'day_price' across the years
fig = px.line(df, x=df['date'], y=df['day_price'])
fig.update_layout(title='Daily Day Price', title_x=0.5, title_y=0.95 ) # Center the title (0.5 means centered)

df2 = df.copy()

# Set the 'date' column as the index
df2.set_index('date', inplace=True)

# Now resample 'day_price' by month and compute the mean
monthly_avg = df2['day_price'].resample('M').mean()

# Create a line plot with monthly averages
fig = px.line(monthly_avg, x=monthly_avg.index, y=monthly_avg, labels={'x': 'Date', 'y': 'Average Day Price'})
fig.update_layout(title='Monthly Average Day Price', title_x=0.5, title_y=0.95 ) # Center the title (0.5 means centered)

# Show the plot
fig.show()

C:\Users\Deninjo\AppData\Local\Temp\ipykernel_13476\694275928.py:7: FutureWarning:

'M' is deprecated and will be removed in a future version, please use 'ME' instead.

px.histogram(df, x=df['day_price'])

import plotly.graph_objects as go
from plotly.subplots import make_subplots


# Create subplots for each numerical column
fig = make_subplots(rows=3, cols=3, subplot_titles=[f'Boxplot for {col}' for col in numericals])

# Loop through each numerical column and create a boxplot
for i, column in enumerate(numericals, 1):
    row = (i - 1) // 3 + 1  # Determine the row index (1 to 3)
    col = (i - 1) % 3 + 1   # Determine the column index (1 to 3)
    
    fig.add_trace(
        go.Box(
            y=df[column],  # Data for the boxplot
            name=column       # Name of the boxplot (will appear in the title)
        ),
        row=row, col=col       # Specify the location of the subplot
    )

# Update the layout to add a title and improve spacing
fig.update_layout(
    title_text="Boxplots for Numerical Columns",  # Title for the entire plot
    title_x=0.5,  # Center the title
    height=900,   # Adjust the height of the plot
    showlegend=False  # Hide the legend (not needed for boxplots)
)

# Show the plot
fig.show()

from scipy.stats import mstats

df3 = df.copy()


# Winsorize the data for each numerical column
for column in numericals:
    df3[column] = mstats.winsorize(df3[column], limits=[0.2, 0.2])  # Replace outliers with 20th and 80th percentiles

# Create subplots for each numerical column
fig = make_subplots(rows=3, cols=3, subplot_titles=[f'Boxplot for {col}' for col in numericals])

# Loop through each numerical column and create a boxplot
for i, column in enumerate(numericals, 1):
    row = (i - 1) // 3 + 1  # Determine the row index (1 to 3)
    col = (i - 1) % 3 + 1   # Determine the column index (1 to 3)
    
    fig.add_trace(
        go.Box(
            y=df3[column],  # Data for the boxplot
            name=column    # Name of the boxplot (will appear in the title)
        ),
        row=row, col=col   # Specify the location of the subplot
    )

# Update the layout to add a title and improve spacing
fig.update_layout(
    title_text="Boxplots for Winsorized Numerical Columns",  # Title for the entire plot
    title_x=0.5,  # Center the title
    height=900,   # Adjust the height of the plot
    showlegend=False  # Hide the legend (not needed for boxplots)
)

# Show the plot
fig.show()

from sklearn.svm import SVR
from sklearn.tree import DecisionTreeRegressor
from sklearn.preprocessing import StandardScaler
from sklearn.neighbors import KNeighborsRegressor
from sklearn.linear_model import LinearRegression, Ridge

from sklearn.metrics import r2_score, mean_absolute_error, mean_squared_error

# Import dataset
df3

# Define X and Y
numericals = ['12m_low', '12m_high', 'day_low', 'day_high','previous','volume']

x = df3[numericals].values
y = df3['day_price'].values

# Split the dataset into training and test set
from sklearn.model_selection import train_test_split

X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=0.2, random_state=42)

# Scale the features to transforms the data into a distribution with:
# A mean of 0
# A standard deviation of 1
scaler = StandardScaler()
X_train_scaled = scaler.fit_transform(X_train)
X_test_scaled = scaler.transform(X_test)

# Initialize models
models = {
    'Linear Regression': LinearRegression(),
    'Ridge Regression': Ridge(alpha=1.0),
    'KNN Regression': KNeighborsRegressor(n_neighbors=5),
    'Decision Tree Regression': DecisionTreeRegressor(),


}

# Evaluate models
results = []

for name, model in models.items():
    model.fit(X_train_scaled, y_train)
    y_pred = model.predict(X_test_scaled)
    mae = mean_absolute_error(y_test, y_pred)
    mse = mean_squared_error(y_test, y_pred)
    rmse = np.sqrt(mse)
    r2 = r2_score(y_test, y_pred)
    results.append([name, mae, mse, rmse, r2])

# Display results
results_df = pd.DataFrame(results, columns=['Model', 'MAE', 'MSE', 'RMSE', 'R2'])
results_df.set_index('Model', inplace=True)
results_df

# Plotly Bar graph
fig = px.bar(
    results_df, 
    x=results_df.index, 
    y=['MAE', 'MSE', 'RMSE', 'R2'], 
    barmode='group', 
    title="Model Evaluation Metrics", 
    labels={"value": "Score", "Model": "Model"},
    height=500
)

# Show the plot
fig.show()

# Find the best model based on the desired metric (R²)
best_model_r2 = results_df['R2'].idxmax()  # Model with highest R²
best_model_mse = results_df['MSE'].idxmin()  # Model with lowest MSE
best_model_mae = results_df['MAE'].idxmin()  # Model with lowest MAE
best_model_rmse = results_df['RMSE'].idxmin()  # Model with lowest RMSE

# Display the best models based on each metric
print(f"Best Model based on R²: {best_model_r2}")
print(f"Best Model based on MSE: {best_model_mse}")
print(f"Best Model based on MAE: {best_model_mae}")
print(f"Best Model based on RMSE: {best_model_rmse}")

Best Model based on R²: Linear Regression
Best Model based on MSE: Linear Regression
Best Model based on MAE: Decision Tree Regression
Best Model based on RMSE: Linear Regression

# Initialize models
lr_model = LinearRegression()
dt_model = DecisionTreeRegressor()

# Fit models to the training data
lr_model.fit(X_train, y_train)
dt_model.fit(X_train, y_train)

# Make predictions on the test set
y_pred_lr = lr_model.predict(X_test)
y_pred_dt = dt_model.predict(X_test)

# Create a figure and axis
plt.figure(figsize=(14, 6))

# Plot predicted vs actual values for Linear Regression
plt.subplot(1, 2, 1)
plt.scatter(y_test, y_pred_lr, color='blue', alpha=0.5)
plt.plot([y_test.min(), y_test.max()], [y_test.min(), y_test.max()], color='red', linestyle='--')  # Line of perfect prediction
plt.title('Linear Regression: Predicted vs Actual')
plt.xlabel('Actual Values')
plt.ylabel('Predicted Values')

# Plot predicted vs actual values for Decision Tree Regressor
plt.subplot(1, 2, 2)
plt.scatter(y_test, y_pred_dt, color='green', alpha=0.5)
plt.plot([y_test.min(), y_test.max()], [y_test.min(), y_test.max()], color='red', linestyle='--')  # Line of perfect prediction
plt.title('Decision Tree Regressor: Predicted vs Actual')
plt.xlabel('Actual Values')
plt.ylabel('Predicted Values')

# Display the plots
plt.tight_layout()
plt.show()

import matplotlib.pyplot as plt

# Calculate residuals for Linear Regression
residuals_lr = y_test - y_pred_lr

# Calculate residuals for Decision Tree Regressor
residuals_dt = y_test - y_pred_dt

# Create residual plot for Linear Regression
plt.figure(figsize=(14, 6))

plt.subplot(1, 2, 1)
plt.scatter(y_pred_lr, residuals_lr, color='blue', alpha=0.5)
plt.axhline(y=0, color='red', linestyle='--')
plt.title('Residual Plot for Linear Regression')
plt.xlabel('Predicted Values (Linear Regression)')
plt.ylabel('Residuals')

# Create residual plot for Decision Tree Regressor
plt.subplot(1, 2, 2)
plt.scatter(y_pred_dt, residuals_dt, color='green', alpha=0.5)
plt.axhline(y=0, color='red', linestyle='--')
plt.title('Residual Plot for Decision Tree Regressor')
plt.xlabel('Predicted Values (Decision Tree Regressor)')
plt.ylabel('Residuals')

# Show the plots
plt.tight_layout()
plt.show()

df3

# Select the features from the i'th row 
features = ['12m_low', '12m_high', 'day_low', 'day_high', 'previous', 'volume']

X_single_row = df3.loc[4, features].values.reshape(1, -1)  # Reshape to 2D array

# Make a prediction using the trained model
y_pred_single = lr_model.predict(X_single_row)

# Get the actual price from the df3['day_price'] column for the same row
y_actual_single = df3.loc[4, 'day_price']

# Output the prediction and actual value
print(f"Predicted value : {y_pred_single[0]}")
print(f"Actual value : {y_actual_single}")

Predicted value : 126.91836872539568
Actual value : 127.0

	DATE	CODE	NAME	12m Low	12m High	Day Low	Day High	Day Price	Previous	Change	Change%	Volume	Adjust
0	1/2/2007	EGAD	Eaagads	22	57	52	52	52	52	-	-	300	26
1	1/2/2007	KAPC	Kapchorua Tea	111	148	100	100	100	98.5	1.5	1.50%	7,800	-
2	1/2/2007	KUKZ	Kakuzi	67.5	89	42.5	45	43.5	42.25	1.25	2.87%	2,700	-
3	1/2/2007	REA	Rea Vipingo Plantations	14.5	23.5	25	26.5	25.5	25.75	-0.25	0.98%	10,500	-
4	1/2/2007	SASN	Sasini Tea and Coffee	10.5	13.6	140	145	140	141	-1	0.71%	70,900	23.33
...	...	...	...	...	...	...	...	...	...	...	...	...	...
10603	12/24/2007	EVRD	Eveready East Africa	1.4	3	7	8.1	7.95	7.65	0.3	3.77%	84,200	-
10604	12/24/2007	MSC	Mumias Sugar	4	7.35	13.7	15	14.8	14.3	0.5	3.38%	2,400,000	-
10605	12/24/2007	UNGA	Unga Group	9.6	16.5	15	15.5	15.45	15.5	-0.05	0.32%	20,300	12.88
10606	12/24/2007	ACCS	AccessKenya Group	3.9	10.1	21.5	24.75	23.25	22.5	0.75	3.23%	1,090,000	-
10607	12/24/2007	^N20I	NSE 20-Share Index	3,396.83	5,030.91	5,444.83	5,444.83	5,444.83	5,339.75	105.08	1.93%	12,530,000	-

	Date	CODE	NAME	12m Low	12m High	Day Low	Day High	Day Price	Previous	Change	Change%	Volume	Adjusted
0	1/2/2008	^N20I	NSE 20-Share Index	3,396.83	5,030.91	5,167.18	5,167.18	5,167.18	5,444.83	-277.65	5.37%	4,790,000	-
1	1/2/2008	ACCS	AccessKenya Group	3.9	10.1	21	25	21.25	23.25	-2	9.41%	538,100	-
2	1/2/2008	UNGA	Unga Group	9.6	16.5	15	15	15	15.45	-0.45	3.00%	25,000	12.5
3	1/2/2008	MSC	Mumias Sugar	4	7.35	13.35	14.2	13.6	14.8	-1.2	8.82%	617,200	-
4	1/2/2008	EVRD	Eveready East Africa	1.4	3	7.2	8	7.7	7.95	-0.25	3.25%	23,900	-
...	...	...	...	...	...	...	...	...	...	...	...	...	...
10578	12/31/2008	C&G	Car and General (K)	21	29	44	44	44	44	-	-	200	29.33
10579	12/31/2008	WTK	Williamson Tea Kenya	180	290	57.5	57.5	57.5	52.5	5	8.70%	1,000	-
10580	12/31/2008	SASN	Sasini Tea and Coffee	10.5	13.6	7	7.35	7	6.7	0.3	4.29%	26,700	-
10581	12/31/2008	REA	Rea Vipingo Plantations	14.5	23.5	13.5	14	13.95	13.65	0.3	2.15%	1,100	-
10582	12/31/2008	KAPC	Kapchorua Tea	111	148	68	68	68	70	-2	2.94%	500	-

	DATE	CODE	NAME	12m Low	12m High	Day Low	Day High	Day Price	Previous	Change	Change%	Volume	Adjust
0	1/2/2009	^N20I	NSE 20-Share Index	3,396.83	5,030.91	3,589.16	3,589.16	3,589.16	3,521.18	67.98	1.89%	3,750,000	-
1	1/2/2009	^NASI	NSE All-Share Index	73.71	121.58	74.2	74.2	74.2	73.37	0.83	1.12%	3,750,000	-
2	1/2/2009	SCOM	Safaricom	3.05	6.5	3.6	3.75	3.6	3.6	-	-	1,220,000	-
3	1/2/2009	ACCS	AccessKenya Group	3.9	10.1	21.25	22.75	22.25	20.75	1.5	6.74%	200,700	-
4	1/2/2009	UNGA	Unga Group	9.6	16.5	12.85	13.6	12.9	13.6	-0.7	5.43%	4,500	-
...	...	...	...	...	...	...	...	...	...	...	...	...	...
14381	12/31/2009	REA	Rea Vipingo Plantations	14.5	23.5	11.5	11.5	11.5	11.25	0.25	2.17%	200	-
14382	12/31/2009	LIMT	Limuru Tea	400	500	305	305	305	305	-	-	0	-
14383	12/31/2009	KUKZ	Kakuzi	67.5	89	31.75	31.75	31.75	31.75	-	-	0	-
14384	12/31/2009	KAPC	Kapchorua Tea	111	148	86	86	86	86	-	-	0	-
14385	12/31/2009	EGAD	Eaagads	22	57	20	20	20	20	-	-	0	-

	DATE	CODE	NAME	12m Low	12m High	Day Low	Day High	Day Price	Previous	Change	Change%	Volume	Adjust
0	1/4/2010	^NASI	NSE All-Share Index	74.68	122.14	71.89	71.89	71.89	71.64	0.25	0.35%	3,730,000	-
1	1/4/2010	^N20I	NSE 20-Share Index	3,443.94	5,030.91	3,261.17	3,261.17	3,261.17	3,247.44	13.73	0.42%	3,730,000	-
2	1/4/2010	KPLC-P7	Kenya Power and Lighting 7%	5.5	5.5	5.5	5.5	5.5	5.5	-	-	0	-
3	1/4/2010	KPLC-P4	Kenya Power and Lighting 4%	7.65	7.65	7.65	7.65	7.65	7.65	-	-	0	-
4	1/4/2010	SCOM	Safaricom	3.1	6.6	4.55	4.65	4.55	4.55	-	-	2,800,000	-
...	...	...	...	...	...	...	...	...	...	...	...	...	...
14947	12/31/2010	REA	Rea Vipingo Plantations	14.5	23.5	15.8	17.5	17.45	17	0.45	2.58%	155,500	-
14948	12/31/2010	LIMT	Limuru Tea	400	500	300	300	300	300	-	-	0	-
14949	12/31/2010	KUKZ	Kakuzi	67.5	89	80	82	81.5	77	4.5	5.52%	67,500	-
14950	12/31/2010	KAPC	Kapchorua Tea	111	148	100	100	100	100	-	-	0	-
14951	12/31/2010	EGAD	Eaagads	22	57	50	50	50	50	-	-	0	-

	DATE	CODE	NAME	12m Low	12m High	Day Low	Day High	Day Price	Previous	Change	Change%	Volume	Adjust
0	1/3/2011	EGAD	Eaagads	22	57	50	53.5	52.5	50	2.5	4.76%	1,200	-
1	1/3/2011	KAPC	Kapchorua Tea	111	148	100	100	100	100	-	-	0	-
2	1/3/2011	KUKZ	Kakuzi	67.5	89	81.5	81.5	81.5	81.5	-	-	0	-
3	1/3/2011	LIMT	Limuru Tea	400	500	300	300	300	300	-	-	0	-
4	1/3/2011	REA	Rea Vipingo Plantations	14.5	23.5	17	18.55	18.45	17.45	1	5.42%	25,000	-
...	...	...	...	...	...	...	...	...	...	...	...	...	...
14926	12/30/2011	SCOM	Safaricom	3.05	6.5	2.9	3.05	2.95	3	-0.05	1.69%	1,590,000	-
14927	12/30/2011	KPLC-P4	Kenya Power and Lighting 4%	7.65	7.65	7.65	7.65	7.65	7.65	-	-	0	-
14928	12/30/2011	KPLC-P7	Kenya Power and Lighting 7%	5.5	5.5	5.5	5.5	5.5	5.5	-	-	0	-
14929	12/30/2011	^N20I	NSE 20-Share Index	3,396.83	5,030.91	3,205.02	3,205.02	3,205.02	3,160.03	44.99	1.40%	4,710,000	-
14930	12/30/2011	^NASI	NSE All-Share Index	73.71	121.58	68.03	68.03	68.03	67.85	0.18	0.26%	4,710,000	-

Stock Market Analysis and Prediction¶

Columns and Description¶

Preparing the dataset¶

Data Cleaning¶

EDA (Exploratory Data Analysis)¶

Handling Outliers¶

Model Training¶

Using the Model¶

	DATE	CODE	NAME	12m Low	12m High	Day Low	Day High	Day Price	Previous	Change	Change%	Volume	Adjust
0	1/3/2012	^NASI	NSE All-Share Index	73.23	119.37	68.38	68.38	68.38	68.03	0.35	0.51%	10,990,000	-
1	1/3/2012	^N20I	NSE 20-Share Index	3,312.56	4,985.91	3,212.86	3,212.86	3,212.86	3,205.02	7.84	0.24%	10,990,000	-
2	1/3/2012	KPLC-P7	Kenya Power and Lighting 7%	5.5	5.5	5.5	5.5	5.5	5.5	-	-	0	-
3	1/3/2012	KPLC-P4	Kenya Power and Lighting 4%	7.65	7.65	7.65	7.65	7.65	7.65	-	-	0	-
4	1/3/2012	SCOM	Safaricom	3.05	6.5	2.9	3	2.95	2.95	-	-	7,410,000	-
...	...	...	...	...	...	...	...	...	...	...	...	...	...
16039	12/31/2012	REA	Rea Vipingo Plantations	14.5	23.5	19	19	19	19	-	-	1,000	-
16040	12/31/2012	LIMT	Limuru Tea	400	500	430	430	430	430	-	-	0	-
16041	12/31/2012	KUKZ	Kakuzi	67.5	89	72	72	72	69	3	4.17%	1,000	-
16042	12/31/2012	KAPC	Kapchorua Tea	111	148	118	118	118	118	-	-	100	-
16043	12/31/2012	EGAD	Eaagads	22	57	25	25	25	25	-	-	0	-

	date	code	name	12m_low	12m_high	day_low	day_high	day_price	previous	change	change%	volume	adjusted
0	2007-01-02	EGAD	Eaagads	22.0	57.0	52.0	52.0	52.0	52.00	NaN	NaN	300	26.00
1	2007-01-02	KAPC	Kapchorua Tea	111.0	148.0	100.0	100.0	100.0	98.50	1.50	0.0150	7800	NaN
2	2007-01-02	KUKZ	Kakuzi	67.5	89.0	42.5	45.0	43.5	42.25	1.25	0.0287	2700	NaN
3	2007-01-02	REA	Rea Vipingo Plantations	14.5	23.5	25.0	26.5	25.5	25.75	-0.25	0.0098	10500	NaN
4	2007-01-02	SASN	Sasini Tea and Coffee	10.5	13.6	140.0	145.0	140.0	141.00	-1.00	0.0071	70900	23.33
...	...	...	...	...	...	...	...	...	...	...	...	...	...
81499	2012-12-31	REA	Rea Vipingo Plantations	14.5	23.5	19.0	19.0	19.0	19.00	NaN	NaN	1000	NaN
81500	2012-12-31	LIMT	Limuru Tea	400.0	500.0	430.0	430.0	430.0	430.00	NaN	NaN	0	NaN
81501	2012-12-31	KUKZ	Kakuzi	67.5	89.0	72.0	72.0	72.0	69.00	3.00	0.0417	1000	NaN
81502	2012-12-31	KAPC	Kapchorua Tea	111.0	148.0	118.0	118.0	118.0	118.00	NaN	NaN	100	NaN
81503	2012-12-31	EGAD	Eaagads	22.0	57.0	25.0	25.0	25.0	25.00	NaN	NaN	0	NaN

	date	code	name	12m_low	12m_high	day_low	day_high	day_price	previous	change	change%	volume	adjusted
15053	2008-05-30	HFCK-R	NaN	11.05	11.05	5.00	20.00	6.55	NaN	6.55	1.0	17200	NaN
22070	2009-02-02	KPLC-P7	Kenya Power and Lighting 7%	5.50	5.50	5.00	5.00	5.00	NaN	5.00	1.0	0	NaN
22079	2009-02-02	CARB	Carbacid Investments	88.50	140.00	137.00	137.00	137.00	NaN	137.00	1.0	0	45.67
22080	2009-02-02	BOC	B.O.C Kenya	90.00	130.00	160.00	160.00	160.00	NaN	160.00	1.0	0	NaN
22099	2009-02-02	UCHM	Uchumi Supermarket	12.70	24.00	14.50	14.50	14.50	NaN	14.50	1.0	0	NaN
44813	2010-08-16	TPSE-R	TPS Eastern Africa (Serena) (Rights)	5.10	5.10	5.00	8.00	5.40	NaN	5.40	1.0	700	NaN
45711	2010-09-07	SCBK-R	Standard Chartered Bank (Rights)	51.00	89.00	70.00	70.00	70.00	NaN	70.00	1.0	500	NaN
49242	2010-12-01	KPLC-R	Kenya Power and Lighting (Rights)	0.55	0.55	4.00	4.50	4.00	NaN	4.00	1.0	11100	NaN
58420	2011-07-14	TCL	Trans-Century	20.00	37.75	50.00	60.00	57.00	NaN	57.00	1.0	1140000	NaN
69653	2012-04-03	KQ-R	Kenya Airways (Rights)	0.05	0.05	0.05	0.05	0.05	NaN	0.05	1.0	228200	NaN
72010	2012-05-30	LKL	Longhorn Kenya	8.50	21.00	15.00	20.00	16.15	NaN	16.15	1.0	4200	NaN
74136	2012-07-18	CIC	CIC Insurance Group	3.05	7.00	4.00	7.00	5.25	NaN	5.25	1.0	1620000	NaN

	date	code	name	12m_low	12m_high	day_low	day_high	day_price	previous	change	change%	volume	adjusted
7359	2007-09-05	OCH-R	NaN	0.60	0.60	0.50	1.00	0.95	19.20	-18.25	19.2105	253000	NaN
7403	2007-09-06	OCH-R	NaN	0.60	0.60	0.45	1.00	0.45	0.95	-0.50	1.1111	511800	NaN
7448	2007-09-07	OCH-R	NaN	0.60	0.60	0.05	0.50	0.20	0.45	-0.25	1.2500	404800	NaN
7492	2007-09-10	OCH-R	NaN	0.60	0.60	0.15	0.70	0.40	0.20	0.20	0.5000	272200	NaN
7537	2007-09-11	OCH-R	NaN	0.60	0.60	0.20	1.00	0.20	0.40	-0.20	1.0000	5170000	NaN
7583	2007-09-12	OCH-R	NaN	0.60	0.60	0.25	1.00	0.55	0.20	0.35	0.6364	329000	NaN
7627	2007-09-13	OCH-R	NaN	0.60	0.60	0.10	1.00	0.40	0.55	-0.15	0.3750	3040000	NaN
7671	2007-09-14	OCH-R	NaN	0.60	0.60	0.40	2.00	0.60	0.40	0.20	0.3333	3650000	NaN
8723	2007-10-22	NIC-R	NaN	5.05	98.00	54.50	55.00	54.50	191.00	-136.50	2.5046	1000	NaN
8764	2007-10-23	NIC-R	NaN	5.05	98.00	90.00	100.00	95.00	54.50	40.50	0.4263	5100	NaN
8807	2007-10-24	NIC-R	NaN	5.05	98.00	95.00	100.00	99.00	95.00	4.00	0.0404	11200	NaN
8850	2007-10-25	NIC-R	NaN	5.05	98.00	95.00	100.00	96.00	99.00	-3.00	0.0313	8500	NaN
8894	2007-10-26	NIC-R	NaN	5.05	98.00	87.00	95.00	89.00	96.00	-7.00	0.0787	4600	NaN
8940	2007-10-29	NIC-R	NaN	5.05	98.00	84.00	90.00	85.00	89.00	-4.00	0.0471	10700	NaN
8985	2007-10-30	NIC-R	NaN	5.05	98.00	85.50	95.00	93.50	85.00	8.50	0.0909	5000	NaN
9029	2007-10-31	NIC-R	NaN	5.05	98.00	93.00	100.00	95.00	93.50	1.50	0.0158	142100	NaN
9071	2007-11-01	NIC-R	NaN	5.05	98.00	90.00	100.00	92.00	95.00	-3.00	0.0326	35500	NaN
9118	2007-11-02	NIC-R	NaN	5.05	98.00	90.00	100.00	98.00	92.00	6.00	0.0612	13000	NaN
9195	2007-11-06	DTK-R	NaN	2.30	50.00	9.00	9.00	9.00	15.70	-6.70	0.7444	4100	NaN
9239	2007-11-07	DTK-R	NaN	2.30	50.00	26.50	26.50	26.50	9.00	17.50	0.6604	16000	NaN
9283	2007-11-08	DTK-R	NaN	2.30	50.00	12.00	26.00	23.50	26.50	-3.00	0.1277	8800	NaN
9325	2007-11-09	DTK-R	NaN	2.30	50.00	12.00	15.00	12.50	23.50	-11.00	0.8800	8600	NaN
9368	2007-11-12	DTK-R	NaN	2.30	50.00	12.00	14.00	12.05	12.50	-0.45	0.0373	64100	NaN
15097	2008-06-03	HFCK-R	NaN	11.05	11.05	8.00	11.00	10.05	6.55	3.50	0.3483	24200	NaN
15141	2008-06-04	HFCK-R	NaN	11.05	11.05	8.00	15.00	11.85	10.05	1.80	0.1519	131900	NaN
15186	2008-06-05	HFCK-R	NaN	11.05	11.05	11.00	15.00	13.05	11.85	1.20	0.0920	5820000	NaN
15230	2008-06-06	HFCK-R	NaN	11.05	11.05	8.00	14.00	9.70	13.05	-3.35	0.3454	17540000	NaN
15272	2008-06-09	HFCK-R	NaN	11.05	11.05	9.00	13.50	9.25	9.70	-0.45	0.0486	162500	NaN
15316	2008-06-10	HFCK-R	NaN	11.05	11.05	9.55	11.00	10.45	9.25	1.20	0.1148	118200	NaN
15359	2008-06-11	HFCK-R	NaN	11.05	11.05	10.00	12.00	10.40	10.45	-0.05	0.0048	137100	NaN
15404	2008-06-12	HFCK-R	NaN	11.05	11.05	10.05	13.25	11.65	10.40	1.25	0.1073	254700	NaN
15448	2008-06-13	HFCK-R	NaN	11.05	11.05	10.00	12.50	11.05	11.65	-0.60	0.0543	342600	NaN
15708	2008-06-23	KCB-R	NaN	0.25	0.25	5.60	6.80	6.60	217.00	-210.40	31.8788	86700	NaN
15753	2008-06-24	KCB-R	NaN	0.25	0.25	5.50	6.80	5.75	6.60	-0.85	0.1478	147700	NaN
15797	2008-06-25	KCB-R	NaN	0.25	0.25	5.50	31.25	9.85	5.75	4.10	0.4162	76200	NaN
15843	2008-06-26	KCB-R	NaN	0.25	0.25	4.35	30.00	6.35	9.85	-3.50	0.5512	93800	NaN
15888	2008-06-27	KCB-R	NaN	0.25	0.25	1.00	34.75	5.25	6.35	-1.10	0.2095	1010000	NaN
15933	2008-06-30	KCB-R	NaN	0.25	0.25	5.00	6.00	5.35	5.25	0.10	0.0187	109100	NaN
15979	2008-07-01	KCB-R	NaN	0.25	0.25	5.05	6.50	5.35	5.35	NaN	NaN	133700	NaN
16026	2008-07-02	KCB-R	NaN	0.25	0.25	5.05	6.00	5.45	5.35	0.10	0.0183	239200	NaN
16074	2008-07-03	KCB-R	NaN	0.25	0.25	4.35	31.25	5.05	5.45	-0.40	0.0792	3790000	NaN
16122	2008-07-04	KCB-R	NaN	0.25	0.25	5.00	5.50	5.25	5.05	0.20	0.0381	165300	NaN
16170	2008-07-07	KCB-R	NaN	0.25	0.25	3.00	5.35	5.15	5.25	-0.10	0.0194	246800	NaN
16218	2008-07-08	KCB-R	NaN	0.25	0.25	4.80	5.30	5.00	5.15	-0.15	0.0300	45720000	NaN
16265	2008-07-09	KCB-R	NaN	0.25	0.25	5.05	6.00	5.10	5.00	0.10	0.0196	13580000	NaN
16311	2008-07-10	KCB-R	NaN	0.25	0.25	5.20	9.60	5.45	5.10	0.35	0.0642	448300	NaN
16357	2008-07-11	KCB-R	NaN	0.25	0.25	5.50	7.00	5.95	5.45	0.50	0.0840	4720000	NaN

	MAE	MSE	RMSE	R2
Model
Linear Regression	0.175972	0.367902	0.606549	0.999823
Ridge Regression	0.176169	0.369004	0.607457	0.999822
KNN Regression	0.242779	0.673791	0.820848	0.999676
Decision Tree Regression	0.174512	0.436381	0.660591	0.999790